Disrupting business with team data science at IBM Insight at World of Watson 2016
Team data science is the core competency for digital disruption in today’s business environment. I participated in a CrowdChat, 13 September 2016, in which subject matter experts and industry influencers discussed this theme. Here are some highlights from what I contributed.
Which C-level executives drive the most disruptive data-driven business apps?
Data-driven digital marketing is table stakes in every industry, hence the chief marketing officer (CMO) is in the forefront. But in terms of fresh disruption, my sense is C-level executives responsible for product development and research and development (R&D) are driving cognitive Internet of Things embedded in every product.
What value can enterprise data science and engineering teams provide in driving business disruption?
Data engineers and data scientists can collect, aggregate, prepare, mine and collaboratively explore all the relevant data within an enterprise data lake. Competitive disruption comes from collaboratively and creatively exploring and testing data-driven alternatives. Within a team, data engineers, data scientists, business analysts, domain experts and developers are that team. And building a great data science center of excellence doesn’t require that the sponsoring executives be top-notch data scientists themselves. But they do need to know how to spot and recruit that talent.
Is a role for so-called citizen data scientists available in disruptive data science?
Self-taught data scientists are fresh blood. To the extent that they engage with teams of more experienced data scientists, data engineers and so on, they can be valuable in driving ideation and exploring new tools. Essentially, self-taught data scientists are in abundance at startups everywhere, given that established data scientists are in short supply and fetch high salaries. See my recent TechTarget column on citizen data scientists in an enterprise context. You can interpret citizen data scientist as a power business analyst who got religion—and education—on data science, machine learning, deep learning, Apache Spark and so on—and not a naive newbie where data-driven analysis is concerned. And you can also interpret citizen data scientist as someone who applies data science skills and tools, however acquired, to humanitarian initiatives—perhaps as a full-time vocation or as a volunteer or on sabbatical from a day job.
How can data-driven applications drive business disruption?
Disrupting the competitive landscape requires insight, agility and speed that your rivals may lack. Smarter data, more sophisticated analytics and empowered decision makers deliver all those aspects. If you can deliver intelligent apps into market that drive next-best actions continuously at all touch points, that’s disruptive. It enables all stakeholders to do much more with much less. In addition, decision automation—powered by predictive models, rules engines and streaming analytics—can overcome that process bottleneck to deliver continuously and consistently relevant insights to every touch point under every scenario. Predictive and prescriptive decision optimization becomes the norm.
Automotive is most ripe for data-driven transformation. It’s already underway: self-driving and semi-autonomous vehicles, Uber and so on. I’m from Detroit, and I can’t help but think the Motor City is becoming a breeding ground of digital geeks.
Healthcare is obviously being transformed inside and out by machine learning, data science and cognitive computing. Continuous self-monitoring and tracking through Internet of Things and wearable devices and prosthetics is also a hugely disruptive trend. And then media and entertainment has been revolutionized inside and out over the past 20 years. Deep learning, streaming analytics, cognitive Internet of Things and mobile data are fundamental disruptors. Data scientists are dominant media and entertanment app developers.
What is the collaboration workflow that is well suited for data science and engineering projects to drive business disruption?
On some level, the well-suited team needs to have an internal competitive dynamic as brilliant team members compete amongst themselves and engage in nonstop ideation, A/B testing, and the like—in other words, a Lennon-McCartney (Harrison, Starr, Martin) dynamic. On another level, the well-suited collaborative dynamic needs a steady stream of exogenous challenges—competitive pressures that compel everybody to focus on common, do-or-die business imperatives. Startups face this pressure constantly. I think the epitome of a data science team for disruptive apps needs a strong mix of fresh blood—that is, millennials—and old blood—people in my generation or slightly younger. Strong experience but also passion and ideation are vital.
What common curriculum best prepares team members to drive development of disruptive data apps?
Though four years old, this blog on data science curriculum still applies full force. For disruptive outcomes, the opimal curriculum is fail fast, which serves as the school of hard knocks for remorsely winnowing the field of nice ideas in data-driven apps to zero in on the crazy ones that actually change the game. Domain knowledge is key to disruption. Data scientists of all stripes should learn the chief business applications of data science in their organizations and in how to work best with subject-domain experts.
Team data science will be a key theme at the upcoming IBM Insight at World of Watson 2016 event in Las Vegas, Nevada, 24–27 October 2016. Here are some relevant sessions that you should check out at the conference. Use the session preview tool to see further details:
- 1079: Apache Spark and IBM Streams Working Together in Streaming Analytics
- 2954: How We Use Data Science Experience to Analyze Data Science Experience
- 3079: How IBM and BlocPower are Modeling Energy Usage with Data Science Experience
- 3288: Drive Real Data into Real Actionable Insight with Machine Learning on Real-Time Big Data
- 3310: Data Science Made Easy with IBM: A User Panel Discussion
- 3416: Data Scientists: Why Organizations Want One, What They Can Do, and How to Spot a Fake
Architecture solutions and patterns
- 2137: Using Open Metadata to Drive Better Data Science Projects
Data integration and governance
- 1651: IBM DataWorks for the Data Scientist: Mingling Data at Scale
- 1652: IBM DataWorks for the Business Analyst: Self-Service Data Blending
- 1106: Use Cases at the Intersection of Apache Spark and Apache Hadoop and z Analytics
- 1479: Sparkified dashDB
- 3516: Accelerate Your Data Science Delivery with Integrated Notebooks and IBM BigInsights
- 3561: A Data Science Introduction for Database Girls and Guys
You’ll benefit from attending IBM Insight at World of Watson 2016, which represents the continuing evolution of the event formerly known as IBM Insight. We look forward to seeing you there.